As circuit patterns for semiconductors become finer, higher levels of resolution are demanded of steppers and scanners that expose these patterns. To satisfy demands for higher resolution, the wavelength of the radiation employed must be reduced, and the numerical aperture (NA) of the optical system must be increased.
Only a few optical materials are adequately transmissive at shorter wavelengths. For wavelengths of 300 nm or less, the only currently available materials that can be used effectively are synthetic fused silica and fluorite.
The Abbe numbers of fused silica and fluorite are not sufficiently different from each other to allow complete correction of chromatic aberration. For this reason, at wavelengths of 300 nm or below, it is extremely difficult to correct chromatic aberration in projection-optical systems comprised solely of standard refractive optical systems.
Fluorite itself suffers from certain disadvantages. The refractive index of fluorite changes relatively rapidly with variations in temperature, and fluorite polishes poorly. Thus, many optical systems do not use fluorite, resulting in systems with lenses of fused silica only. Such all-silica systems exhibit uncorrectable chromatic aberration.
Purely reflective optical systems avoid chromatic aberration, but such systems tend to be excessively large, and to require one or more aspheric reflecting surfaces. The production of (large) precision aspheric surfaces is extremely difficult.
As a result, various technologies making use of "catadioptric" optical systems (i.e., optical systems in which refractive elements are combined with a reflective elements) have been proposed for reduction projection-optical systems. Among these have been several that propose the formation of an intermediate image one or more times within the optical system.
Previously proposed reduction projection-optical systems which form only one intermediate image are disclosed in Japanese laid-open patent documents 5-25170 (1993), 63-163319 (1988), 4-234722 (1992), and in U.S. Pat. No. 4,779,966. Among these proposed systems, only those disclosed in Japanese laid-open patent document 4-234722 and U.S. Pat. No. 4,779,966 use just one concave mirror.
Japanese laid-open patent document 4-234722 and U.S. Pat. No. 4,779,966 disclose catadioptric optical projection systems comprising a concave mirror and a double-pass lens group. Incident light propagates through the double-pass lens group in a first direction, strikes the concave mirror, and then propagates, as reflected light, back through the double-pass lens group in a second direction opposite to the first direction. Because the double-pass lens groups of Japanese laid-open patent document 4-234722 and U.S. Pat. No. 4,779,966 use only concave lenses and thus have negative power, the light entering the concave mirror is dispersed, requiring a relatively large-diameter concave mirror.
The double-pass lens group of Japanese laid-open patent document 4-234722 (1992) is completely symmetric, which reduces aberrations to an extreme degree, significantly reducing the aberration correction burden for the downstream refractive optical system. However, the completely symmetric configuration also reduces the distance between the intermediate image and the nearest optical element to such a degree that use of a beam-splitter is necessitated to effectively redirect the reflected light while allowing passage of the incident light.
The optical system disclosed in U.S. Pat. No. 4,779,966 comprises a concave mirror in a second imaging system that images an intermediate image onto the wafer. To provide adequate image brightness in this configuration, divergent light enters the concave mirror, requiring a relatively large-diameter mirror.
In optical systems utilizing several mirrors, it is possible to reduce the number of refractive lenses, but other problems arise.
In order to obtain adequate depth of focus with improved resolution, phase-shift reticles are often used. To most effectively use a phase-shift reticle, the ratio .sigma. between the illuminating optical system NA and the imaging optical system NA should be variable. An aperture stop can be installed in the imaging system to provide or increase this variability. But, in a catadioptric imaging system, as, for example, in U.S. Pat. No. 4,779,966, there is often no location for an effective aperture stop.
In catadioptric optical systems in which a double-pass lens system is employed in a demagnifying portion of the optical system, the demagnification reduces the allowable distance between the reflecting element and the wafer, so that few lenses can be placed in the optical path between the reflective element and the wafer. This necessarily limits the numerical aperture (NA), and thus the maximum brightness, of the optical system. Even if it were possible to realize an optical system with a high NA, many optical elements would have to be placed along a limited optical path length, so that the distance between the wafer and the tip surface of the object lens (i.e., the working distance WD) would be undesirably short.
In conventional catadioptric optical systems, the optical path must be eccentric over at least a portion of its length. The adjustment procedure for the eccentric sections of such optical systems is difficult and makes the realization of precision systems essentially impossible.
The applicant has previously proposed a dual-imaging optical system which is designed with a first imaging system comprising a two-way optical system having a concave mirror and a double-pass lens group that allows light both incident to, and reflected from, the concave mirror to pass through the lens group. An intermediate image is formed by the first imaging system and an image of the intermediate image is formed by a second imaging system. A reflecting surface is provided to direct the light flux from the first imaging system toward the second imaging system.
This dual-imaging optical system allows a smaller diameter concave mirror, and provides an effective aperture-stop placement position, allowing a variable ratio .sigma. based on the illuminating optical system NA and the imaging system NA for use with phase-shift reticles for resolution enhancement. It also allows for sufficient optical system brightness and an optical system where the working distance WD, the distance between the wafer and the nearest surface of the object imaging system, can be relatively long. It also makes the adjustment of the eccentric section of the optical system easy, enabling the practical realization of a precision optical system.
While this dual-imaging optical system has many superior features, attempts to reduce the size of the optical system while maintaining image-forming performance results in increased distortion. That is, the optical system is not symmetric, so even if other aberrations are corrected, distortion will remain.
Also, when trying to correct distortion, astigmatism correction may be affected, and it is well known that it is extremely difficult to correct both types of aberration at the same time.
It is desirable to leave other types of well-corrected aberration as-is, and correct only the distortion or astigmatism aberration, especially the higher-order distortion.
In the manufacturing of high-precision optical systems, variance from product to product inevitably arises due to manufacturing tolerances. This variance results in different aberration levels for each optical system produced. Such manufacturing-error-induced aberrations are normally corrected by adjusting sections of the optical system. However, when there is asymmetric aberration of differing amounts across the image surface due to manufacturing tolerances, or when the generated aberration amounts are too great, it is often impossible to fully correct the system for manufacturing tolerances solely by adjusting sections of the optical system. In this case, corrections can sometimes be made by inserting an aspheric, aberration-correcting plate near the final focused image. While such a correcting plate is effective in correcting "angle-of-view" aberrations (such as distortion and/or astigmatism) when placed as close as possible to the image surface, in practice, the presence of other adjusting devices or measuring equipment near the image surface normally requires that such plates be placed a sufficient distance away from the image surface such that other types of aberration (related to aperture) are also affected. This complicates the correction process.